Site-Specific Response to Nitrogen Fertilization in the Argentinean Pampas.

The objective was to evaluate a set of variable rate nitrogen applications and comparing the crop response in homogeneous zones. The experiment was conducted in 92 hectares field in a randomized complete block design arranged as three plots. 20, 138 and 256 kg ha^-1 of nitrogen were randomly assigned to each plot. We selected managements zones as a blocking factor. Blocks were assigned from site properties (apparent electrical conductivity (ECa), effective depth and elevation) and yield (NDVI 2004-2010). Two zoning methods were used: first, principal component spatially restricted analysis (MULTISPATI-PCA) and then, cluster analysis with non-hierarchical algorithm fuzzy k-means in the first three spatial principal component. Four management zones were identified as a function of minimum value achieved by NCE and FPI index. Yields obtained under different fertilization treatments in interaction with management zones were evaluated using ANOVA with block random effect and error spatial correlation. Significant and positive fertilization response was obtained between the three management zones (one high-yielding and other two with relatively low- yielding), although differences between doses were not significant. High-yielding management zone was characterized by deep soil and low ECa. The response was different between low-yielding management zones with positive fertilization effect. One of them had lesser effective depth, lesser nitrate contain and low organic matter content, and other ones had higher clay content. Results indicated that the methods were adequate to delineate homogeneous zones in terms of site specific nitrogen fertilization. It was concluded that to properly delineate management zones, it is important to identify those soil variables responsible for production performance in the Argentina Pampas.